Clostridium botulinum is an anaerobically growing, sporulating bacterium, which produces a highly toxic protein. This so-called botulinum toxin is the cause of botulism, a food poisoning, which without the use of intensive care measures can lead to the death of botulism patients. Seven serotypes are distinguished (type A-G, shortly termed BoNT/A, BoNT/B, etc.) that have a similar amino acid sequence, but induce a different antibody response. The toxins (hereinafter also referred to as neurotoxins and botulinum toxins) consist of two functional chains, the light (˜50 kDa) and the heavy chain (˜100 kDa), which are generated by proteolytic cleavage of the single-chain precursor protein. Other strains do not possess the corresponding protease, therefore the cleavage into the chains takes place in the gastrointestinal tract of the patients (e.g. by trypsin). In the double-chain form the subunits (i.e., the heavy and the light chain) are interconnected via disulfide bridges (for example, in addition there exists an intramolecular disulfide bridge in BoNT/A, i.e., between two cysteine residues of the heavy chain).
Under acidic conditions in vivo the pure neurotoxins do not exist in free form, but form complexes with other clostridial proteins, the so-called (Clostridium botulinum) toxin complexes. Different proteins, inter alia with hemagglutinating properties, are involved in these complexes. The composition of the complexes is different between serotypes. The integration into the complex protects the neurotoxin during gastrointestinal passage. These other clostridial proteins (the complexing and complex proteins, respectively) possibly play also a role in the resorption of the neurotoxin. Thus, the incorporation in the complex causes the neurotoxin to be orally bioavailable and to thus constitute a food poison. The target location of the neurotoxins is at the motor endplate, where the muscle is activated by the nerve. The motoneuron releases acetylcholine for activation of the muscle. This release is inhibited by botulinum toxin. The inhibitory effect takes place in 3 sequential steps: binding, translocation, proteolysis. The heavy chain of botulinum toxin binds highly specific to the motoneuron and is subsequently taken up into the nerve cell by endocytosis. Upstream of the binding domain, which is located at the C-terminal end of the heavy chain, there is the translocation domain in the N-terminal portion of the heavy chain, which transports or, rather, facilitates the translocation of the light chain into the cytosol by an as yet unknown mechanism. In the cytosol the light chain becomes active as a protease cleaving highly specific so-called SNARE proteins. The proteolytic specificity of the individual botulinum toxin types is summarized in Table 1. These SNARE proteins are responsible for the fusion of the acetylcholine-loaded secretory vesicles with the cell membrane of the motoneuron. The proteolytic cleavage of one of these SNARE proteins inhibits the formation of a fusion complex and thus further release of acetylcholine. The affected muscle is no longer activated. Previously hyperactive muscles become paralyzed.
TABLE 1SNARE proteinBotulinum ToxinSubstrate of proteaseCleavage site in SNARETypeactivitysequence from ratType ASNAP 25EANQ197 RATKType BVAMP 2GASQ76 FETSType CSyntaxinDTKK254 AVKYSNAP 25ANQR198 ATKType DVAMP 2RDQK61 LSEDType ESNAP 25QIDR180 IMEKType FVAMP 2ERDQ60 KLSEType GVAMP 2ETSA83 AKLK
This mechanism of action is taken advantage of in the therapy of a multitude of muscle disorders and spasms, respectively, characterized by an uncontrolled release of acetylcholine (e.g., blepharospasm, torticollis, spasticity) Extremely low amounts of the neurotoxin (in the pg to ng range) are injected in the hyperactive muscle for therapy of dystonia. The neurotoxin diffuses to the motor endplate and reaches the cytosol of the neuron to inhibit the acetylcholine release there. The muscle is paralyzed after 1-2 days.
Various facial wrinkles are formed by cramping of muscles lying beneath the skin, thus also through uncontrolled release of acetylcholine. Botulinum toxins find cosmetic utilization in this context: wrinkles will be removed for about 3 months through injection of extremely low amounts of botulinum toxin.
At present four preparations containing botulinum toxins have received drug-regulatory approval: Botox® (Allergan), Xeomin® (Merz), Dysport® (Ipsen), and NeuroBloc® (Solstice Neurosciences). Botox®, Xeomin®, and Dysport® are lyophilisates of botulinum toxin type A (as complex, neurotoxin and complex, respectively), Botox® and Xeomin® with 100 units per injection vial each, Dysport® with 500 units. NeuroBloc® contains botulinum toxin type B (as complex) with 5,000 and 10,000 units, respectively, in liquid formulation.
Except for NeuroBloc the preparations are available as lyophylisates, which are reconstituted with physiological saline and are injected in the respective muscles in matched doses depending on preparation and indication. The treated muscle will be paralyzed within 48 h. The effect lasts about 3 month, thereafter a further injection must be carried out, if the muscle should remain paralyzed further, i.e. the dystonia is to be treated. Up to now it has not unambiguously elucidated, which processes control the decrease of the effect. As long as the light chain is active as protease, the appropriate SNARE protein is cleaved (e.g., SNAP 25 through the light chain of neurotoxin type A).
Accordingly, the fusion of the secretory vesicles with the plasma membrane and thereby the release of acetylcholine will be inhibited under these conditions, the muscle remains paralyzed. If it were possible to maintain the protease activity of the light chain for an extended time period in the cell, then the duration of action of an appropriate drug would be extended also
In contrast to many low molecular active substances active protein substances are characterized by a significantly lower stability. The half life (HL) of some active protein substances in the circulating blood amounts to only a few minutes, so that the (therapeutic) duration of action is strongly restricted and injection must be repeated in short intervals. The HL can be extended, if one is successful in protecting the protein against degradation and elimination processes. One theoretically possible way exists especially for eukaryotic proteins in a higher glycosylation (more carbohydrate moieties) and in adapting the carbohydrate structures to the structures of human glycoproteins, respectively. Another path that has been taken in a series of approved active substances is the coupling of the protein with polyethylene glycol (PEG). PEG can be covalently bonded to the residues of various amino acids, e.g., to lysine (amino function) or cysteine residues (SH function). PEG enhances the molecular weight of the protein without creating immunogenic structures that induce the generation of antibodies to the active substance. To the contrary: the PEGylation reduces the immunogenicity of the active substance. The protein is eliminated more slowly by the increase of the molecular weight and a significant increase in HL is achieved. For maintaining a certain required serum level, the drug has to be injected less often.
PEGylated active protein substances are already processed in some approved drugs (see Table 2). The employment of the partly small proteins (e.g., interferon α 2a:Mr=19.3 kDa) in the original form, i.e. not modified, has shown that the proteins are very rapidly eliminated from the serum. The PEGylation gave rise to a markedly increased molecular weight and thus to a substantially longer half life in the serum. Thus, for example, the serum half life for interferon α 2a is 9 h; PEGylation with a 40 kDa PEG chain drastically increases the molecular weight and extends the half life from 9 to 72 h.
TABLE 2TradenameStarting compoundCoupling of PEGPegasysinterferon α 2abranched PEG-N-hydroxysuccinimide;Coupling to 4 lysine residuesNeulastaG-CSFPEG-aldehyde;Coupling to N-terminal methioninePeglutroninterferon α 2bSuccinimidyl carbonate-PEG;Coupling to histidine and lysineresiduesSomavestgrowth hormone4-6 PEG;antagonistCoupling to lysine residues and N-terminusOncasparAsparaginaseN-hydroxysuccinimide activated PEG
However the linkage with one or more PEG chains is subject to restrictions:                1. Preferably the PEG chain diminishes the biological activity of the modified protein (in comparison with the unmodified native protein) not at all or only slightly (in accordance with the invention it is understood that slightly diminished biological activity of the modified protein corresponds to at least 20%, preferably to 30-40% or 50-70% or even to 75-95% of the biological activity of the unmodified native protein). A diminished activity is tolerable in many cases: e.g. the antiviral activity of PEGylated interferon is 25-35% of the non-PEGylated interferon α 2b. PEGylated interferon α 2a even possesses only 1-7% of the activity of the non-PEGylated form.        2. As a multitude of therapeutically employed proteins deploy their activity through the binding to a specific receptor, preferably the PEGylation does not affect, or only slightly affects, the interaction with the receptor (e.g., the interaction can be affected directly by steric hindrance at the binding domain or by alterations of the spatial structure of the protein that have an effect on the binding domain and hence on binding).        3. When the pharmacological effect of the therapeutic protein is (also) mediated through an enzymatic activity (as for instance with asparaginase), preferably the enzymatic activity is not, or only slightly, reduced through the PEGylation.        
Preferably the PEGylation of botulinum toxin accomplishes these three criteria. At the same time the modification of the botulinum toxin with PEG preferably influences neither (a) the binding domain of the heavy chain nor (b) the enzymatic activity of the light chain, i.e., the PEG chain preferably does not inhibit the interaction of the catalytic domain from the light chain with the substrate (SNARE Protein). In contrast to other proteases, that cleave short peptides, botulinum toxins require longer peptides as substrates. For instance, a peptide which serves as a substrate for botulinum toxin type B preferably has a sequence of about 40 amino acid residues of the SNARE protein VAMP 2. Peptides with shorter SNARE sequences will also be cleaved, but with substantially lower efficiency. The cleavage domain of the light chain of the botulinum toxin, which has a length comparable to the recognition sequence of about 40 amino acid residues, is preferably not affected by the PEG chain. Moreover, it has to be considered, that besides the cleavage domain responsible for the direct contact of the substrate (SNARE protein and peptide with the SNARE sequence of about 40 amino acid residues, respectively) with the light chain, additional contact sites with sequences on the light chain located distantly to the catalytic domain are needed for optimal activity of botulinum toxin. It has been demonstrated that five additional contact sites for its substrate SNAP 25 are localized on the light chain of botulinum toxin type A: 4 α exosites (AS 102-113, 310-321, 335-348, 351-358) and one β exosite (AS 242-259). Preferably, the contact is not or only marginally impeded through a conjugation of the light chain with PEG. Moreover, the C-terminal part of the heavy chain, the translocation domain, must be operable, i.e., it must ensure that the light chain is transported from the endosomes into the cytosol. This transport process that is absolutely necessary for the action can also be inhibited through the steric hindrance of a PEGylated light chain especially as the translocation domain possibly forms a pore in the endosomal membrane through which a “bulky” PEGylated light chain might not be channeled through.
Coupling of PEG to botulinum toxin is reported in a U.S. patent application (2002/0197278). The coupling serves to diminish the antigenicity and immunogenicity, respectively as well as to enhance the molecular weight for reducing the diffusion. For the selection of the appropriate sites (antigenic determinants) and amino acid residues, respectively, for the PEGylation, reference is being had to the paper of Bavari et al. (Vaccine 16: 1850-1856, 1998). In this paper sequences of the botulinum toxin heavy chain that induce neutralizing antibodies are presented. In the aforementioned patent application it is only stated that (1) the PEGylation should be carried out at, respectively close to one site or at, respectively, close to the sites, which act(s) as an important epitope(s), but which are remote from the catalytic domain (i.e. remote from the light chain) and that (2) PEG may be conjugated to the free terminal carboxy or amino groups or at the amino groups of lysine side chains. (3) As additional alternative for the insertion of PEG into the toxin it is suggested to use the SH groups of naturally occurring or specially inserted cysteine residues; however, the paper advises against this alternative (3), as disulfide bridges between the heavy and the light chain of the botulinum toxin play a role in the spatial configuration of the molecule. There is no example given that discloses the structure of the PEGylated neurotoxin or that discloses on which amino acid residue(s) a PEG molecule of a certain length was attached.
In a further patent application (WO 02/40506) relating to the change in stability, the insertion, the modification or the removal of sites for the in vivo glycosylation, in vivo phosphorylation and primarily the in vivo myristoylation in the botulinum toxin are suggested in order to optionally either enhance or decrease the stability of the botulinum toxin. A whole series of potential modification sites are specified which are located at a significant distance to the N- and C-terminal ends of the neurotoxin light chain. Additional sequences are to be inserted into the polypeptide chain, where carbohydrate chains or phosphate and myristoyl moieties, respectively, are coupled at the light chain by cellular enzymes. Information regarding an accordingly modified neurotoxin or its preparation is however missing.
In a further U.S. patent application (2003/0027752) a peptide residue with a so-called leucine motif (e.g., XEXXXLL) is inserted into the neurotoxin or into the light chain in order to increase the stability of the light chain within the nerve cell. The configuration of the light chain with this motif ensures that it is localized in the vicinity of its substrate at the membrane. Moreover, a so-called “tyrosine based motif” (YXXHy, Y=tyrosine, Hy=hydrophobic amino acid) is set forth that, after insertion in the light chain, is to enhance its persistency. Finally this patent application suggests a modified botulinum toxin type A, in which the light chain is mutated (alanine to leucine at the positions 427 and 428).
In view of the above described prior art it was an object of the inventors to provide an additional or precisely described form of stabilization for any type of botulinum toxin preferably, however, for type A, B, and C1. Along with this the object of the inventors was to provide stable variants/analogs of the natural botulinum toxins which in comparison to the respective unmodified botulinum toxins have an increased in vivo stability. This means, firstly, that the biological activity (according to the invention biological activity is defined as total activity comprising the enzymatic/catalytic activity of the light chain as well as the required neurotoxin binding to the target cell and the translocation of the light chain into the target cell) of the botulinum toxin variant/analog shall be at most marginally (according to the above definition), and preferably not at all, decreased and, secondly, that, in spite of its modification, the light chain is translocated to its site of action, the cytosol of the motoneuron.
In contrast to the already aforementioned US 20020197278 the objective forming the basis of the present application does not aim to block antigenic determinants, to decrease the antigenicity of the toxins or to restrict their diffusion away from the injection site.
The inventors of the present application surprisingly found, that the light chain of botulinum toxins can be specifically PEGylated at its N-terminus via insertion of at least one cysteine residue without simultaneously impairing or even inhibiting the biological activity (according to the definition given above) of the botulinum toxins. Such a PEGylated botulinum toxin is characterized by a surprisingly higher in vivo stability (significant increase of HL and therewith an extended (pharmacological) duration of action).
The (therapeutic) duration of action of the natural botulinum toxins in the patient depends on the serotype. Botulinum toxin type A is characterized by the longest duration of action of about 3 month. The duration of action of botulinum toxin type C is of similar length as that of type A, whereas botulinum toxin type B has a shorter duration of action. The effect of botulinum toxins type E and F lasts only about 2 weeks in each case. The short duration of action of these two types does not allow their clinical application for the treatment of dystonia. The present invention allows (1) the clinical application of all botulinum toxins, even those having so far short-term activity, and (2) a more advantageous therapy with the already therapeutically utilized type A and B toxins, as these need not be administered every three month, but e.g., only every six month.